As a method for full-text retrieval that comprehensively retrieves a targeted character/letter string from voluminous documents, there are known a grep type retrieval (sequential retrieval) and an index type retrieval. The grep type retrieval is typified by a grep command in UNIX (registered trademark), whilst the index type retrieval is typified by an inverted index, and is an indexing retrieval that makes use of an index.
The former retrieval, viz., the grep type retrieval, sequentially scans files as the subjects of retrieval. Hence, the response performance is lowered in case the number of files being retrieved has increased drastically. Thus, if desired to implement the retrieval function for the voluminous information, the index type retrieval is used. Non-Patent Document 1 is directed to explanation of the principle of the index type retrieval that makes use of the above mentioned inverted file.
Recently, the above mentioned index type retrieval has been ushered extensively into a field different from a search engine of the Internet. Patent Document 1 has disclosed an e-mail system having the function of the full-text retrieval by an index.
In the above-mentioned e-mail system, it is basically necessary that indices are prepared and held separately on the user-by-user basis. The reason is that, if a plurality of users make use of a sole index, it becomes necessary to provide a scheme for managing access control, thus possibly deteriorating the performance. On the other hand, if the weighted representation (ranking) of the results of retrieval is to be realized in accordance with, for example, a TF-IDF method, it becomes necessary to provide a separate scheme of not allowing a keyword appearing in a mail of a given user to be reflected in the weighting for another user (see Non-Patent Document 2).
As typical of the techniques of updating the indices of a database, there are the following techniques. Viz., Patent Document 2 has disclosed a document storage retrieval system in which an index for retrieval is prepared not synchronously with respect to document storage but asynchronously. Patent Document 3 has disclosed a data base system of performing retrieval with the aid of an index, in which an index is generated or dropped based on the access frequency of each attribute (key) of data to be retrieved.    [Patent Document 1] JP Patent Kokai Publicaion No. JP-P2001-036568A    [Patent Document 2] JP Patent Kokai Publicaion No. JP-A-3-276368    [Patent Document 3] JP Patent Kokai Publicaion No. JP-A-6-215037    [Non-Patent Document 1] Kenji Kita, Kazuhiko Tsuda and Masamiki Shishibori, ‘Information Retrieval Algorithm’, Kyoritsu Shuppan Publishing Co. Ltd., January 1, 2002, pp. 6 and pp.160 to 179    [Non-Patent Document 2] Stefan Buettcher and Charles L. A. Clarke, “A Security Model for Full-text File System Search in Multi-User Environments”, In proceedings of the 4th Conference on USENIX Conference on File and Storage Technologies (FAST05)- Volume 4 (San Francisco, Calif. Dec. 13-16, 2005), USENIX Association Berkeley Calif. 13-13